Adaptive, arousal-related adjustments of perceptual biases optimize perception in a dynamic environment

Abstract

Prior expectations can be used to improve perceptual judgments about ambiguous stimuli. However, little is known about if and how these improvements are maintained in dynamic environments in which the quality of appropriate priors changes from one stimulus to the next. Using a novel sound-localization task, we show that changes in stimulus predictability lead to arousal-mediated adjustments in the magnitude of prior-driven biases that optimize perceptual judgments about each stimulus. These adjustments depend on task-dependent changes in the relevance and reliability of prior expectations, which subjects update using both normative and idiosyncratic principles. The resulting variations in biases across task conditions and individuals are reflected in modulations of pupil diameter, such that larger stimulus-evoked pupil responses correspond to smaller biases. These results suggest a critical role for the arousal system in adjusting the strength of perceptual biases with respect to inferred environmental dynamics to optimize perceptual judgements.

The complexity dividend: when sophisticated inference matters

“…Consistent with this idea, learning rates with more relevance to an ongoing estimate of choice values have been shown to explain more variance in fMRI signals (Meder et al, 2017). This weighting process may be regulated by noradrenergic, cholinergic, and dopaminergic neuromodulatory systems, each of which has been linked to adaptive inference via pupillometry and other measures (Nassar et al, 2012;Krishnamurthy et al, 2016;Krugel et al, 2009;Aston-Jones & Cohen, 2005;Joshi et al, 2016).…”

Perceptual Behavior Depends Differently on Pupil-Linked Arousal and Heartbeat Dynamics-Linked Arousal in Rats Performing Tactile Discrimination Tasks

“…Previous studies have demonstrated that non-luminance mediated fluctuations of pupil size was able to track rapid changes in cortical state (Reimer et al, 2014(Reimer et al, , 2016McGinley et al, 2015a;Vinck et al, 2015), and therefore pupil size has been widely considered as a peripheral index of arousal (Nassar et al, 2012;Murphy et al, 2014;Ebitz and Platt, 2015;Lee and Margolis, 2016;Krishnamurthy et al, 2017;Urai et al, 2017;Schriver et al, 2018Schriver et al, , 2020de Gee et al, 2020). It has long been postulated that the LC is the primary brain structure mediating task evoked pupil dilations (Aston-Jones and Cohen, 2005).…”

Eye pupil signals information gain

“…In de Gee et al [18,22], pupil responses in detection and 2-alternative forced choice tasks were shown to be inversely proportional to the probability of the choice and hence to the KL divergence between prior and posterior: in conservative participants (biased towards NO), YES choices led to larger responses, while the opposite tended to be found in more liberal participants (biased towards YES). Pupil responses were also shown to vary as a function of the influence of the prior on perceptual decisions in de Gee et al [22] and Krishnamurthy et al [21]: when prior beliefs have less weight (because of better control or attentional allocation or because of low prior reliability), more information is extracted from the sensory stimulus, KL divergence is larger, and the pupil dilates more. Along the same line, when the occurrence of surprising outcomes suggests the task structure may have changed, pupil dilation is even larger [10,21,26].…”

“…Besides the well-known response of pupillary muscles to light, which narrows the range of light intensity reaching the retina and optimizing its information capacity [1], pupil size varies also as a function of a wealth of cognitive phenomena, including mental effort [2][3][4][5], surprise [6][7][8][9][10][11][12][13][14][15], emotion [16], decision processes [17][18][19][20], decision biases [19,21,22], value beliefs [23][24][25], volatility (unexpected uncertainty; [10,[26][27][28]), exploitation/exploration trade-off [29,30], attention [31][32][33][34][35][36], uncertainty [12,19,21,23,25,37,38], confidence [39], response to reward [40], learning rate…”